# Import Data Below
medical_student <- read_csv(file = "DataMedTeach.csv", show_col_types = FALSE)
medical_student
## # A tibble: 886 × 20
##       id   age  year   sex glang  part   job stud_h health  psyt  jspe qcae_cog
##    <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>  <dbl>  <dbl> <dbl> <dbl>    <dbl>
##  1     2    18     1     1   120     1     0     56      3     0    88       62
##  2     4    26     4     1     1     1     0     20      4     0   109       55
##  3     9    21     3     2     1     0     0     36      3     0   106       64
##  4    10    21     2     2     1     0     1     51      5     0   101       52
##  5    13    21     3     1     1     1     0     22      4     0   102       58
##  6    14    26     5     2     1     1     1     10      2     0   102       48
##  7    17    23     5     2     1     1     0     15      3     0   117       58
##  8    21    23     4     1     1     1     1      8      4     0   118       65
##  9    23    23     4     2     1     1     1     20      2     0   118       69
## 10    24    22     2     2     1     1     0     20      5     0   108       56
## # … with 876 more rows, and 8 more variables: qcae_aff <dbl>, amsp <dbl>,
## #   erec_mean <dbl>, cesd <dbl>, stai_t <dbl>, mbi_ex <dbl>, mbi_cy <dbl>,
## #   mbi_ea <dbl>

Creator: Betissa Kouassi-Brou

Q1: What role does the year of study of the participant play into the level of burnout?

library(ggplot2)
ggplot(medical_student, aes(x = factor(year), y = mbi_ex))+
  geom_boxplot() +
  labs(x = "Year of Study", y = "Burnout Level")

burnout_by_year <- aggregate(mbi_ex ~ year, data = medical_student, FUN = mean)

burnout_table <- kbl(burnout_by_year, align = "c") %>%
  kable_classic(full_width = F) %>%
  kable_styling("striped", font_size = 14)

burnout_table
year mbi_ex
1 17.67755
2 18.46667
3 17.88811
4 16.59350
5 15.32283
6 14.02655

Q2: Is there a significant linear relationship between burnout levels and hours spent studying among medical students, and if so, what is the strength of this relationship?

plot(medical_student$stud_h, medical_student$mbi_ex, pch = 19, col = "lightblue",
     xlab = "Student hours", ylab = "MBI exhaustion")

abline(lm(medical_student$mbi_ex ~ medical_student$stud_h), col = "red", lwd = 3)

cor_val <- round(cor(medical_student$mbi_ex, medical_student$stud_h), 2)
text(x = 25, y = 95, paste("Correlation:", cor_val))

mtext(paste("Correlation:", cor_val), side = 1, line = 2)

Interpreter: Yuvraj Jain

Q1: Is there a disparity in the Job Satisfaction Score between genders?

I start off by creating a density plot to visualize the data at a first glance, and see whether this question is worth investigating. JSPE scores refer to job satisfaction scores calculated on the JSPE (Jefferson Scale of Physician Empathy) scale through 20 items, each answered on 7-point Likert scale (1 ¼ Strongly Disagree, 7 ¼ Strongly Agree).

# Create a box plot to compare the job satisfaction score between genders
medical_student_i_1 <- medical_student

ggplot(medical_student_i_1, aes(x = jspe, fill = factor(sex))) +
  geom_density(alpha = 0.5) +
  scale_fill_manual(values = c("#00BFC4", "#F8766D", "#FFFFB0"), name = "Gender",
                    labels = c("Male", "Female", "Non-Binary")) +
  theme_minimal() +
  labs(x = "JSPE Score", y = "Density",
       title = "Distribution of JSPE Scores by Gender") +
  theme(plot.title = element_text(hjust = 0.5, size = 14, face = "bold"),
        axis.text = element_text(size = 12),
        axis.title = element_text(size = 12, face = "bold"),
        legend.position = "top",
        legend.text = element_text(size = 12),
        legend.title = element_text(size = 12, face = "bold"),
        panel.border = element_rect(color="gray30", fill=NA, linewidth=1))

sum(medical_student$sex == 3)
## [1] 5

In order to perform a t-test, we can only do the comparison between two groups. Since I initially based off the question to compare the job satisfaction scores between males and females, and there is only 5 responses from non-binary genders which is not enough to make any statistical statements, the subsequent code for this section excludes non-binary people; however, doing a comparison with non-binary people could also yield interesting results, provided sufficient data.

# Create a box plot to compare the job satisfaction score between males & females.
medical_student_i_2 <- subset(medical_student, sex != 3)

ggplot(medical_student_i_2, aes(x = jspe, fill = factor(sex))) +
  geom_density(alpha = 0.5) +
  scale_fill_manual(values = c("#00BFC4", "#F8766D"), name = "Gender",
                    labels = c("Male", "Female")) +
  theme_minimal() +
  labs(x = "JSPE Score", y = "Density",
       title = "Distribution of JSPE Scores by Gender") +
  theme(plot.title = element_text(hjust = 0.5, size = 14, face = "bold"),
        axis.text = element_text(size = 12),
        axis.title = element_text(size = 12, face = "bold"),
        legend.position = "top",
        legend.text = element_text(size = 12),
        legend.title = element_text(size = 12, face = "bold"),
        panel.border = element_rect(color="gray30", fill=NA, linewidth=1))

# Conduct two-sample t-test
t.test(jspe ~ sex, data = medical_student_i_2)
## 
##  Welch Two Sample t-test
## 
## data:  jspe by sex
## t = -3.3288, df = 490.27, p-value = 0.0009377
## alternative hypothesis: true difference in means between group 1 and group 2 is not equal to 0
## 95 percent confidence interval:
##  -3.4648501 -0.8927977
## sample estimates:
## mean in group 1 mean in group 2 
##        104.8327        107.0116

In summary, the test results indicate that there is a statistically significant difference (p-value ~ 0.001) in mean job satisfaction score between Females and Males, with Female participants reporting higher levels of job satisfaction than Male participants on average.

Orator 1: Anjali Mehta

Q1: How does language spoken by medical students relate to their mental health and burnout levels?

language_codes <- c("1"="French", "15"="German", "20"="English", "37"="Arab", "51"="Basque", "52"="Bulgarian", "53"="Catalan", "54"="Chinese", "59"="Korean", "60"="Croatian", "62"="Danish", "63"="Spanish", "82"="Estonian", "83"="Finnish", "84"="Galician", "85"="Greek", "86"="Hebrew", "87"="Hindi", "88"="Hungarian", "89"="Indonesian", "90"="Italian", "92"="Japanese", "93"="Kazakh", "94"="Latvian", "95"="Lithuanian", "96"="Malay", "98"="Dutch", "100"="Norwegian", "101"="Polish", "102"="Portuguese", "104"="Romanian", "106"="Russian", "108"="Serbian", "112"="Slovak", "113"="Slovenian", "114"="Swedish", "116"="Czech", "117"="Thai", "118"="Turkish", "119"="Ukrainian", "120"="Vietnamese", "121"="Other")

medical_6 <- medical_student %>%
  select(glang, psyt, stai_t, mbi_ex, mbi_cy, mbi_ea) %>%
  group_by(glang) %>%
  summarise(
    'Psychological_Distress_Score' = mean( psyt ),
    'Anxiety_Inventory'= mean(stai_t),
    'Exhaustion_Burnout'= mean(mbi_ex),
    'Cynicism_Burnout' = mean(mbi_cy),
    'Efficacy_Burnout' = mean(mbi_ea)
  ) %>%
  mutate('Language_Spoken' = language_codes[as.character(glang)]) %>%
   select('Language_Spoken', everything(), -glang)

#table
medical_6 %>%
  kbl() %>%
  kable_styling()
Language_Spoken Psychological_Distress_Score Anxiety_Inventory Exhaustion_Burnout Cynicism_Burnout Efficacy_Burnout
French 0.2329149 42.48675 16.82008 10.019526 24.16457
German 0.1935484 42.87097 16.38710 9.741936 24.77419
English 0.2272727 40.54545 16.45455 10.636364 25.09091
Arab 0.3333333 55.66667 22.33333 9.666667 22.66667
Chinese 0.0000000 49.00000 26.00000 14.000000 21.00000
Croatian 0.0000000 54.66667 16.00000 7.666667 25.66667
Spanish 0.4000000 42.40000 16.20000 9.400000 24.40000
Italian 0.1777778 43.24444 16.66667 10.111111 24.22222
Japanese 0.0000000 45.00000 13.00000 16.000000 21.00000
Lithuanian 0.0000000 41.00000 15.00000 15.000000 29.00000
Dutch 0.0000000 49.00000 18.00000 8.000000 21.00000
Portuguese 0.1481481 48.85185 18.29630 10.925926 24.44444
Romanian 0.2500000 38.75000 15.50000 8.000000 26.25000
Russian 0.1666667 45.50000 17.16667 10.500000 25.83333
Serbian 0.0000000 32.00000 8.00000 5.000000 27.00000
Swedish 0.0000000 49.00000 24.00000 15.000000 20.00000
Turkish 0.5000000 47.00000 22.00000 11.500000 25.00000
Vietnamese 0.0000000 53.50000 17.00000 12.500000 21.00000
Other 0.2307692 47.84615 18.23077 11.153846 22.61538 
# Psychological Distress Scores
ggplot(medical_6[, c(1, 2)], aes(x = Language_Spoken, y = Psychological_Distress_Score, fill = Language_Spoken)) +
  geom_bar(stat = "identity") +
  labs(x = "Language Spoken", y = "Psychological Distress Score", title = "Psychological Distress Scores by Language Spoken") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1),
        legend.position = "none")  # rotate x-axis labels for better readability

# Anxiety Inventory
ggplot(medical_6[, c(1, 3)], aes(x = Language_Spoken, y = Anxiety_Inventory, fill = Language_Spoken)) +
  geom_bar(stat = "identity") +
  labs(x = "Language Spoken", y = "Anxiety Inventory", title = "Anxiety Inventory by Language Spoken") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1),
        legend.position = "none")  # rotate x-axis labels for better readability

# Exhaustion Burnout 
ggplot(medical_6[, c(1, 4)], aes(x = Language_Spoken, y = Exhaustion_Burnout, fill = Language_Spoken)) +
  geom_bar(stat = "identity") +
  labs(x = "Language Spoken", y = "Exhaustion Burnout", title = "Exhaustion Burnout Scores by Language Spoken") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1),
        legend.position = "none")  # rotate x-axis labels for better readability

# Cynicism Burnout 
ggplot(medical_6[, c(1, 5)], aes(x = Language_Spoken, y = Cynicism_Burnout, fill = Language_Spoken)) +
  geom_bar(stat = "identity") +
  labs(x = "Language Spoken", y = "Cynicism_Burnout", title = "Cynicism_Burnout Scores by Language Spoken") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1),
        legend.position = "none")  # rotate x-axis labels for better readability

#Efficacy Burnout
ggplot(medical_6[, c(1, 6)], aes(x = Language_Spoken, y = Efficacy_Burnout, fill = Language_Spoken)) +
  geom_bar(stat = "identity") +
  labs(x = "Language Spoken", y = "Efficacy_Burnout", title = "Cynicism_Burnout Scores by Language Spoken") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1),
        legend.position = "none")  # rotate x-axis labels for better readability

correlations <- medical_student %>%
  select(glang, psyt, stai_t, mbi_ex, mbi_cy, mbi_ea)%>%
  cor()

# Print correlation matrix
correlations%>%
  kbl() %>%
  kable_styling()
glang psyt stai_t mbi_ex mbi_cy mbi_ea
glang 1.0000000 -0.0428184 0.0918513 0.0380150 0.0369215 -0.0016969
psyt -0.0428184 1.0000000 0.2932823 0.1772418 0.1457021 -0.1625439
stai_t 0.0918513 0.2932823 1.0000000 0.5304859 0.3318845 -0.4625348
mbi_ex 0.0380150 0.1772418 0.5304859 1.0000000 0.5051998 -0.4808207
mbi_cy 0.0369215 0.1457021 0.3318845 0.5051998 1.0000000 -0.5659386
mbi_ea -0.0016969 -0.1625439 -0.4625348 -0.4808207 -0.5659386 1.0000000

Q2: What is the relationship between academic efficacy and satisfaction with health?

medical_7 <- medical_student %>%
  select(health, mbi_ea)

#table
medical_7[1:10,] %>%
  arrange(desc(health))%>%
  kbl() %>%
  kable_styling()
health mbi_ea
5 21
5 23
4 26
4 23
4 27
3 20
3 23
3 16
2 18
2 22 
# Calculate means by health and academic efficacy
health_efficacy <- medical_7 %>%
  group_by(health) %>%
  summarize(mean_academic_efficacy = mean(mbi_ea))

#table
health_efficacy %>%
  kbl() %>%
  kable_styling()
health mean_academic_efficacy
1 24.62162
2 21.60920
3 22.83824
4 24.24129
5 25.91964 
# Plot relationship between health and academic efficacy
ggplot(health_efficacy, aes(x = health, y = mean_academic_efficacy)) +
  geom_bar(stat = "identity") +
  labs(title = "Relationship between Academic Efficacy and Health", x = "Health", y = "Mean Academic Efficacy Score") +
  theme_bw()

# Calculate correlation between health and academic efficacy
cor(health_efficacy$health, health_efficacy$mean_academic_efficacy)%>%
  kbl() %>%
  kable_styling()
x
0.4967542

Orator 2: Yutong Wu

Q1: Does the job statisfaction score directly impact participants’ health with respect to health levels, anxiety inventory scale, academic motivation scores, and hours of study per week?

ggplot(data = medical_student, aes(x = jspe, y = health, color = "Health")) + 
  geom_point(size = 3) +
  labs(x = "JSPE Scores", y = "Health Scores", 
       title = "Relationship between JSPE Scores and Health") +
  theme(plot.title = element_text(size = 14, face = "bold")) +
  scale_color_manual(values = c("#0072B2"))

ggplot(data = medical_student, aes(x = jspe, y = stai_t, color = "State Anxiety")) + 
  geom_point(size = 3) +
  labs(x = "JSPE Scores", y = "State Anxiety Scores", 
       title = "Relationship between JSPE Scores and State Anxiety") +
  theme(plot.title = element_text(size = 14, face = "bold")) +
  scale_color_manual(values = c("#E69F00"))

ggplot(data = medical_student, aes(x = jspe, y = amsp, color = "AMSP")) + 
  geom_point(size = 3) +
  labs(x = "JSPE Scores", y = "AMSP Scores", 
       title = "Relationship between JSPE Scores and AMSP") +
  theme(plot.title = element_text(size = 14, face = "bold")) +
  scale_color_manual(values = c("#009E73"))

ggplot(data = medical_student, aes(x = jspe, y = stud_h, color = "Study Habits")) + 
  geom_point(size = 3) +
  labs(x = "JSPE Scores", y = "Study Habits Scores", 
       title = "Relationship between JSPE Scores and Study Habits") +
  theme(plot.title = element_text(size = 14, face = "bold")) +
  scale_color_manual(values = c("#CC79A7"))

Q2: What is the relation between anxiety inventory scale and academic motivation scores? What are the factors causing the difference, i.e anxiety impacting academic motivation in any way?

ggplot(data = medical_student, aes(x = factor(amsp), y = stai_t)) + 
  geom_boxplot( color = "#0072B2", alpha = 0.8, size = 0.8) +
  labs(x = "AMSP Scores", y = "State Anxiety Scores", 
       title = "Relationship between AMSP Scores and State Anxiety")

We can tell from the graph that there is a relative negative trend in the relation between anxiety level and academic motivation. The students with higher academic motivation have a relative lower anxiety level.

Deliverer: Jillian Myler

Q1: :How does Partnership Status (part) contribute to self evaluated well being, including MBI emotional exhaustion and satisfaction with health(health)?

partnership_status=
  medical_student %>% 
  select(c(part))


partnership_status
## # A tibble: 886 × 1
##     part
##    <dbl>
##  1     1
##  2     1
##  3     0
##  4     0
##  5     1
##  6     1
##  7     1
##  8     1
##  9     1
## 10     1
## # … with 876 more rows
ps<-ggplot(partnership_status, aes(x=part)) + geom_histogram(fill='blue',color='black', bins =2)
  
ps + theme_minimal()

part_plot<-ps +  xlab("% in partnership") + ylab("Density") + 
  ggtitle("Distribution of Med Students in Partnership")  +
  scale_fill_manual(values=c("red","green")) +
  guides(fill=guide_legend(title="part"))+ 
  geom_vline(xintercept = .50) + 
  geom_hline(yintercept = 1.46)+
  theme_minimal()

part_plot + theme_minimal()

health_sat =
  medical_student %>%
  select(c(health))
hs<-ggplot(health_sat, aes(x=health)) + geom_histogram(fill='blue',color='black', bins=5)
  
hs + theme_minimal()

plot1=
  ggplot(medical_student, aes(x=health, fill =as.factor(part)))+
  geom_histogram(bins=5)+
  scale_fill_manual(values=c('red','green'))+
  facet_wrap(~part)+
  labs(fill = "Partnership Status")

plot1 + theme_minimal()

Q2: What is the relationship between emotional exhaustion and cynicism? Does this relationship relate in a particular manner to academic efficacy?

ex_cynic_figure=
ggplot(medical_student,aes(x=mbi_ex,y=mbi_cy))+
  geom_point()

ex_cynic_figure

cynic_figure_line <- ex_cynic_figure + geom_abline(aes(intercept = mean(mbi_ex) - mean(mbi_cy),
                           slope = 1),
                       linetype = 2, color = "red")
cynic_figure_line

another_one <- cynic_figure_line+ geom_smooth(method = "lm")

another_one
## `geom_smooth()` using formula = 'y ~ x'

compare_to_academics<-
  another_one+
  facet_wrap(~mbi_ea)

compare_to_academics
## `geom_smooth()` using formula = 'y ~ x'
## Warning in qt((1 - level)/2, df): NaNs produced
## Warning in max(ids, na.rm = TRUE): no non-missing arguments to max; returning
## -Inf

Follow-up Questions

New Questions Based Off Initial Investigation

  • Q1: How do mental health metrics vary between different geographic regions?
  • Q2: What other variables in our data correlate with efficacy? What is the secret to becoming an academic weapon?
  • Q3: Why is burnout level decreasing as we go through the years?
  • Q4: How do the different scores measure up if we divide them based on genders? Is there a statistically significant difference between the scores of male and female distributions?

Investigation of Follow-up Questions

Our group decided to investigate Q2 and Q4 in further detail.

Follow-up Question 2: What other variables in our data correlate with efficacy? What is the secret to becoming an academic weapon?

From the correlation plot that we made for one of the initial questions, we had notices that efficacy was weakly positively correlated with stud_h (study hours), health (self-reported health score), jspe (job satisfaction score), qcae_cog (Cognitive empathy measured on the QCAE scale), and amsp (Assessment of Motor and Process Skills). We also saw a weak negative correlation with psyt (Had a psychotherapy test in the last year) and qcae_aff (Affective Empathy score measured on the QCAE scale). However, the most noteworthy observations were the strong negative correlations that efficay had with anxiety, depression, exhaustion, and cynicism. (Using the same terminologies from Interpreter’s section Q2).

Exploring the stronger negative correlations:

library(ggplot2)
library(gridExtra)
## 
## Attaching package: 'gridExtra'
## The following object is masked from 'package:dplyr':
## 
##     combine
data <- medical_student_i_3 

linear_regression <- function(data, x, y) {
  OUTPUT <- ggplot(data, aes(x, y)) + 
    geom_point() +
    geom_smooth(method = "lm", formula = 'y ~ x', se = FALSE) +
    theme_minimal() 
  return (OUTPUT)
}

depression <- linear_regression(data, data$depression, data$efficacy) +
  ggtitle("Relation b/w Depression & Efficacy") + 
  labs(x = "Depression", y = "Efficacy")

anxiety <- linear_regression(data, data$anxiety, data$efficacy) +
  ggtitle("Relation b/w Anxiety & Efficacy") + 
  labs(x = "Anxiety", y = "Efficacy")

exhaustion <- linear_regression(data, data$exhaustion, data$efficacy) +
  ggtitle("Relation b/w Exhaustion & Efficacy") + 
  labs(x = "Exhaustion", y = "Efficacy") 
  
cynicism <- linear_regression(data, data$cynicism, data$efficacy) +
  ggtitle("Relation b/w Cynicism & Efficacy") + 
    labs(x = "Cynicism", y = "Efficacy") 



grid.arrange(depression, anxiety, exhaustion, cynicism, ncol=2)

Exploring the weakly positive correlations:

studyhours <- linear_regression(data, data$stud_h, data$efficacy) +
  ggtitle("Relation b/w Study Hours & Efficacy") + 
  theme(plot.title = element_text(hjust = 0.5, size = 8, face = "bold")) +
  labs(x = "Study Hours", y = "Efficacy") 

jobsatisfaction <- linear_regression(data, data$jspe, data$efficacy) +
  ggtitle("Relation b/w Job Satisfaction Score & Efficacy") + 
  theme(plot.title = element_text(hjust = 0.5, size = 8, face = "bold")) +
  labs(x = "Job Satisfaction Score", y = "Efficacy")

health<-ggplot(data, aes(x = as.factor(health), y = efficacy)) +
  geom_boxplot() +
  ggtitle("Relation b/w Health & Efficacy") +
  theme(plot.title = element_text(hjust = 0.5, size = 8, face = "bold"))+
  labs(x = "Health", y = "Efficacy")+
  theme_minimal()

cogemp <- linear_regression(data, data$qcae_cog, data$efficacy) +
  ggtitle("Relation b/w Cognitive Empathy & Efficacy") +
  theme(plot.title = element_text(hjust = 0.5, size = 8, face = "bold")) +
  labs(x = "Cognitive Empathy Score", y = "Efficacy")

academicmotiv <- linear_regression(data, data$amsp, data$efficacy) +
  ggtitle("Relation b/w Academic Motivation & Efficacy") +
  theme(plot.title = element_text(hjust = 0.5, size = 8, face = "bold")) +
  labs(x = "Academic Motivation", y = "Efficacy")


grid.arrange(studyhours, jobsatisfaction, health, cogemp, academicmotiv, ncol = 3)

Follow-up Quesion 4: How do the different scores measure up if we divide them based on genders? Is there a statistically significant difference between the scores of males and femal

#turn sex into a factor to do the plot color division
x<-as.factor(data$sex)

dep.g<-ggplot(data, aes(x = age, y = depression, color= x)) +
  geom_point() +
  geom_smooth(method = lm, se = FALSE, aes(color = x)) +
  scale_color_manual(values = c("1" = "#F8766D", "2" = "#00BFC4"), 
                     breaks = c(1, 2), 
                     labels = c("Male", "Female")) +
  labs(color = "Gender", fill = "Gender") +
  theme_minimal()

cyn.g <- ggplot(data, aes(x = age, y = cynicism, color = x)) +
  geom_point() +
  geom_smooth(method = lm, se = FALSE, aes(color = x)) +
  scale_color_manual(values = c("1" = "#F8766D", "2" = "#00BFC4"), 
                     breaks = c(1, 2), 
                     labels = c("Male", "Female")) +
  labs(color = "Gender", fill = "Gender") +
  theme_minimal()


exh.g<-ggplot(data, aes(x = age, y = exhaustion, color= x))+
  geom_point() +
  geom_smooth(method = lm, se = FALSE, aes(color = x)) +
  scale_color_manual(values = c("1" = "#F8766D", "2" = "#00BFC4"), 
                     breaks = c(1, 2), 
                     labels = c("Male", "Female")) +
  labs(color = "Gender", fill = "Gender") +
  theme_minimal()

anx.g<-ggplot(data, aes(x = age, y = anxiety, color= x))+
  geom_point() +
  geom_smooth(method = lm, se = FALSE, aes(color = x)) +
  scale_color_manual(values = c("1" = "#F8766D", "2" = "#00BFC4"), 
                     breaks = c(1, 2), 
                     labels = c("Male", "Female")) +
  labs(color = "Gender", fill = "Gender") +
  theme_minimal()

grid.arrange(dep.g,anx.g,cyn.g,exh.g, ncol=2)
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'
## `geom_smooth()` using formula = 'y ~ x'

Running the T-tests:

For the following T-tests, males are group 1 and females are group 2. There were only 5 non-binary people, which is not enough data to perform any tests and come up with reasonable statistically significant conclusions.

medical.student.followup.1 <- medical_student_i_2

# Conduct two-sample t-test for depression in males vs females.
t.test(cesd ~ sex, data = medical.student.followup.1)
## 
##  Welch Two Sample t-test
## 
## data:  cesd by sex
## t = -7.785, df = 624.6, p-value = 2.906e-14
## alternative hypothesis: true difference in means between group 1 and group 2 is not equal to 0
## 95 percent confidence interval:
##  -7.397363 -4.417146
## sample estimates:
## mean in group 1 mean in group 2 
##        14.00364        19.91089
# Conduct two-sample t-test for anxiety in males vs females.
t.test(stai_t ~ sex, data = medical.student.followup.1)
## 
##  Welch Two Sample t-test
## 
## data:  stai_t by sex
## t = -8.1083, df = 542.81, p-value = 3.432e-15
## alternative hypothesis: true difference in means between group 1 and group 2 is not equal to 0
## 95 percent confidence interval:
##  -8.387794 -5.116257
## sample estimates:
## mean in group 1 mean in group 2 
##        38.27273        45.02475
# Conduct two-sample t-test for Exhaustion in Males vs Females
t.test(mbi_ex ~ sex, data = medical.student.followup.1)
## 
##  Welch Two Sample t-test
## 
## data:  mbi_ex by sex
## t = -4.8453, df = 519.11, p-value = 1.671e-06
## alternative hypothesis: true difference in means between group 1 and group 2 is not equal to 0
## 95 percent confidence interval:
##  -2.584502 -1.093325
## sample estimates:
## mean in group 1 mean in group 2 
##        15.61818        17.45710
# Conduct two-sample t-test
t.test(mbi_cy ~ sex, data = medical.student.followup.1)
## 
##  Welch Two Sample t-test
## 
## data:  mbi_cy by sex
## t = -0.33477, df = 528.33, p-value = 0.7379
## alternative hypothesis: true difference in means between group 1 and group 2 is not equal to 0
## 95 percent confidence interval:
##  -0.7662659  0.5431276
## sample estimates:
## mean in group 1 mean in group 2 
##        9.989091       10.100660
# Conduct two-sample t-test
t.test(mbi_ea ~ sex, data = medical.student.followup.1)
## 
##  Welch Two Sample t-test
## 
## data:  mbi_ea by sex
## t = 0.90525, df = 503.26, p-value = 0.3658
## alternative hypothesis: true difference in means between group 1 and group 2 is not equal to 0
## 95 percent confidence interval:
##  -0.3627671  0.9827131
## sample estimates:
## mean in group 1 mean in group 2 
##        24.44364        24.13366

Summary

GIVE A 2 PARAGRAPH SUMMARY.

The initial investigation of the questions revealed interesting results and insights. The dataset was primarily composed of self-reported survey responses, but it also included scores generated from established inventory scales such as MBI and STAI. Most of the columns in the dataset were numerical but represented categorical groups. For instance, sex was represented by 0 for males, 1 for females, and 2 for non-binary. Similarly, other variables like glang, part, job, health, and psyt were also represented by numerical values. Surprisingly, the burnout levels decreased for upperclassmen, which was unexpected as we had anticipated more burnout with longer time spent in medical school. Furthermore, women reported higher job satisfaction scores than men, which was a positive surprise considering the historical gender disparities in pay. Interestingly, partnership status did not affect students’ satisfaction with their health. Instead, it appeared that more people were in partnerships, but their distribution in health satisfaction was similar. Job satisfaction showed a weak correlation with other variables such as health, anxiety, academic motivation, and study habits. The right-skewed trend in job satisfaction indicated that higher job satisfaction led to higher values for all remaining predictors. This led to a question about how these metrics would vary across geographic regions. The investigation into the effects of other variables on academic efficacy also showed similar trends. The decrease in burnout levels as students progressed in medical school led us to question why this was the case. Finally, the question about gender disparities in job satisfaction led to further investigation into gender differences across other metrics and whether any of these differences were statistically significant.

After analyzing the data, we found the second and fourth follow-up questions to be the most promising. As students, we understand that academic efficacy is crucial for future academic or career pursuits, so we decided to investigate the variables that influence it. Additionally, we recognized that the gender wage gap is not the only difference between males and females in medical school, so we explored which metrics differed significantly between genders and how they evolved over time. For the second follow-up, we referred to the correlation plot to study the negative correlations between efficacy and depression, anxiety, exhaustion, and cynicism, creating scatterplots with the line of best fit to examine these correlations. We also investigated weak positive correlations between efficacy and study hours, job satisfaction, health, cognitive empathy, and academic motivation. To explore the fourth follow-up, we created scatterplots to visually represent the relationship between gender and age, depression, anxiety, cynicism, and exhaustion, enabling us to observe performance differences while accounting for age. We used T-tests to discover that there is a statistically significant difference in depression, exhaustion, and anxiety means between males and females but not in efficacy and cynicism. Finally, we discovered that as people age, they are less depressed and anxious, possibly due to a more stable mental state. However, we found that cynicism in females increases with age and plan to investigate this phenomenon further.